Neoplasia is characterized by deregulated cell growth and division. Inevitably, molecular pathways controlling cell growth must interact with those regulating cell division. It was not until very recently, however, that experimental evidence became available to bring such connection to light. Cyclin A was found in association with the adenovirus oncoprotein EIA in virally transformed cells (Giordona et al. Cell 58:981 (1989); and Pines et al. Nature 346:760 (1990)). In an early hepatocellular carcinoma, the human cyclin A gene was found to be the integration site of a fragment of the hepatitis B virus, which leads to activation of cyclin A transcription and a chimeric viral cyclin A protein that is not degradable in vitro (Wang et al. Nature 343:555 (1990)). The cell-cycle gene implicated most strongly in oncogenesis thus far is the human cyclin D1. It was originally isolated through genetic complementation of yeast G.sub.1 cyclin deficient strains (Xiong et al. Cell 65:691(1991); and Lew et al. Cell 66:1197 (1991)), as cellular genes whose transcription is stimulated by CSF-1 in murine macrophages (Matsushine et al. Cell 65:701 (1991)) and in the putative oncogene PRAD1 rearranged in parathyroid tumors (Montokura et al. Nature 350:512 (1991). Two additional human D-type cyclins, cyclins D2 and D3, were subsequently identified using PCR and low-stringency hybridiazation techniques (Inaba et al. Genomics 13:565 (1992); and Xiong et al. Genomics 13:575 (1992)). Cyclin D1 is genetically linked to the bcl-1 oncogene, a locus activated by translocation to an immunoglobulin gene enhancer in some B-cell lymphomas and leukemias, and located at a site of gene amplification in 15-20% of human breast cancers and 25-48% of squamous cell cancers of head and neck origin.
However, the creation of a mutant onocogene is only one of the requirements needed for tumor formation; tumorigenesis appears to also require the additional inactivation of a second class of critical genes: the "anti-oncogenes" or "tumor-suppressing genes." In their natural state these genes act to suppress cell proliferation. Damage to such genes leads to a loss of this suppression, and thereby results in tumorigenesis. Thus, the deregulation of cell growth may be mediated by either the activation of oncogenes or the inactivation of tumor-suppressing genes (Weinberg, R. A., (Sept 1988) Scientific Amer. pp 44-51).
Oncogenes and tumor-suppressing genes have a basic distinguished feature. The oncogenes identified thus far have arisen only in somatic cells, and thus have been incapable of transmitting their effects to the germ line of the host animal. In contrast, mutations in tumor-suppressing genes can be identified in germ line cells, and are thus transmissible to an animal's progeny.
The classic example of a hereditary cancer is retinoblastomas in children. The incidence of the retinoblastomas is determined by a tumor suppressor gene, the retinoblastoma (RB) gene (Weinberg, R. A., (Sept 1988) Scientific Amer.pp 44-51; Hansen et al. (1988) Trends Genet 4:125-128). Individuals born with a lesion in one of the RB alleles are predisposed to early childhood development of retinoblastomas. Inactivation or mutation of the second RB allele in one of the somatic cells of these susceptible individuals appears to be the molecular event that leads to tumor formation (Caveneee et al. (1983) Nature 305:799-784; Friend et al. (1987) PNAS 84:9059-9063).
The RB tumor-suppressing gene has been localized onto human chromosome 13. The mutation may be readily transmitted through the germ line of afflicted individuals (Cavenee, et al. (1986) New Engl. J. Med 314:1201-1207). Individuals who have mutations in only one of the two naturally present alleles of this tumor-suppressing gene are predisposed to retinoblastoma. Inactivation of the second of the two alleles is, however, required for tumorigenesis (Knudson (1971) PNAS 68:820-823).
A second tumor-suppressing gene is the p53 gene (Green (1989) Cell 56:1-3; Mowat et al (1985 Nature 314:633-636). The protein encoded by the p53 gene is a nuclear protein that forms a stable complex with both the SV40 large T antigen and the adenovirus E1B 55 kd protein. The p53 gene product may be inactivated by binding to these proteins.
Based on cause and effect analysis of p53 mutants, the functional role of p53 as a "cell-cycle checkpoint", particularly with respect to controlling progression of a cell from G1 phase into S phase, has implicated p53 as able to directly or indirectly affect cycle cyle machinery. The first firm evidence for a specific biochemical link between p53 and the cell-cycle comes a finding that p53 apparently regulates expression of a second protein, p21, which inhibits cyclin-dependent kinases (CDKs) needed to drive cells through the cell-cycle, e.g. from G1 into S phase. For example, it has been demonstrated that non-viral transformation, such as resulting at least in part from a mutation of deletion of of the p53 tumor suppressor, can result in loss of p21 from cyclin/CDK complexes. As described Xiong et al. (1993) Nature 366:701-704, induction of p21 in response to p53 represents a plausible mechanism for effecting cell-cycle arrest in response to DNA damage, and loss of p53 may deregulate growth by loss of the p21 cell-cycle inhibitor.
The role of RB as a tumor-suppressor protein in cell-cycle control is believed to be similiar to that of p53. However, whereas p53 is generally believed to be responsive to such indigenous environmental cues as DNA damage, the RB protein is apparently involved in coordinating cell growth with exogenous stimulus that normally persuade a cell to cease proliferating, such as diffusible growth inhibitors. In normal cells, RB is expressed throughout the cell cycle but exists in multiple phosphorylated forms that are specfic for certain phases of the cycle. The more highly phosphorylated forms are found during S and G.sub.2 /M, whereas the underphosphorylated forms are the primary species seen in G.sub.1 and in the growth arrested state. Base on these observations, it has been argued that if RB is to have a regulatory (suppressive) activity in the cell-cycle, this activity must be regulated at the post-translational level. Accordingly, underphosphorylated RB would be the form with growth-suppressive acitivty, since this form is prevalent in G.sub.1 and growth arrested cells.
To this end, it is noted that various paracrine growth inhibitors, such as members of the TGF-p family, prevent phosphorylation of RB and arrest cells in late G.sub.1. Current models suggest that during G1, cyclin dependent kinases and particularly cyclin D-associated kinases, CDK4 and CDK6, phosphorylate the product of the retinoblastoma susceptibility gene, RB, and thus release cells from its growth inhibitory eflects. TGF-.beta. treatment causes accumulation of RB in the under-phosphorylated state and expression of RB-inactivating viral oncoproteins prevent TGF-.beta. induced cell cycle arrest. In similar fashion, other related differentiation factors, such as activin, induce accumulation of unphosphorylated RB that is correlated with arrest in G.sub.1 phase.
Recently, it has been demonstrated that the RB protein is a phosphorylation substrate for both CDK4 and CDK6 (Serano et al. (1993) Nature 366:704-707; Kato et al. (1993) Genes Dev 7:331-342; and Meyerson et al. (1994) Mol Cell Biol 14:2077-2086). However, prior to the present discovery, there was little information concerning the manner by which CDK phosphorylation of RB was negatively regulated.